Research Journal of Applied Sciences, Engineering and Technology 4(1): 66-74, 2012
ISSN: 2040-7467
© Maxwell Scientific Organization, 2012
Submitted: September 23, 2011
Accepted: October 28, 2011
Published: January 01, 2012
Determining the Critical Success Factors in ERP Systems Implementations
with Fuzzy Cognitive Mapping: The Case of Turkey
M.S. ¤lkay, 1A.¤. Özdemir, 2G. Seçme and 2N.Y. Seçme
Department of Business, Faculty of Economics and Administrative Sciences,
Erciyes University, Kayseri 38039, Turkey
2
Department of Business, Faculty of Economics and Administrative Sciences,
NevÕehir University, NevÕehir, 50300, Turkey
1
1
Abstract: The purpose of this study is to determine the factors affecting ERP implementation success by using
fuzzy cognitive mapping method. Fuzzy Cognitive Mapping (FCM) is used to determine critical success factors
affecting ERP implementation. FCM is based on graph theory, which was formulized by mathematician Euler
in 1736. Fuzz cognitive method is very useful method for exploring the subjective factors of subject and the
relationships between them. In this study we draw 16 fuzzy cognitive maps with interviewers whose are the
owner or manager of enterprises or ERP team managers of the ERP teams. By the constructive analysis of the
individual cognitive maps, “giving ERP education to team or end users”, “Conformity and reliability of data”
and “Support of top level management” variables are the most central variables. By the neural network analysis
we find that if the ERP education increases, raises; “meeting the business needs of ERP system”, “the efficiency
of project plan”, “project management and time schedule” variables increases too. And according to the second
scenario if top level managers support the use of ERP system increases, raises; “meeting the business needs of
ERP system”, “giving ERP education to team and last users”, “self-sacrifice of project team”, “existence
responsible persons from each department in the project team” variables are increases too. Also, this paper
represents a first attempt for using fuzzy cognitive mapping to determine the factors, which affect the success
of ERP application.
Key words: Critic success factors, ERP, fuzzy cognitive mapping, neural networks, simulation
approach of these studies as following: Traditionally,
scientists are interested in real, objective world that events
could be observed and measured. But, the decision
making processes of the human being did not take place
in this objective world. Always, human decision processes
take place in subjective world of decision makers.
Cognitive mapping opens a window to this subjective
world. By this study, Critical Success Factors of ERP
implementation process will be determined by using fuzzy
cognitive mapping.
INTRODUCTION
Integration of information and communication
technologies throughout the organization and linking all
business units to each other allow managers and
information users of a company to access the needed
information in a timely fashion in order to make right
decisions. Currently, a popular approach to the
development of integrated enterprise-wide system is the
implementation of an Enterprise Resource Planning (ERP)
system (Beheshti, 2006). ERP integrates key business
functions and provides a view of the happenings in the
areas of human resources, finance, accounting,
manufacturing, sales and marketing, procurement,
distribution and etc. of the company to its users.
In this study, cognitive mappings refer to the causal
definitions that could be affected by ERP implementation
success, which are determined by top level managers or
ERP project coordinators of the companies located in
Kayseri, Turkey, by using ERP. The basic assumption of
this approach is that there are internal cognitive models of
the people relating to own observed world (Bauer, 1975).
Fuzzy cognitive mapping has been used in various
studies. Klein and Cooper (1982) summarized the new
LITERATURE REVIEW
Overview of the research on ERP implementation
success factors: ERP system implementation is a process
of great complexity, with a great many conditions and
factors potentially influencing the implementation. These
conditions could have positive effect on the outcome of
ERP project, while their absence could generate problems
during implementation (Soja, 2006). So the causes of
implementation problems or failures need to be
understood and solutions leading to success need to be
found (Calisir and Calisir, 2004).
Corresponding Author: M.S. ¤lkay, Faculty of Economics and Administrative Sciences, Department of Business, Erciyes
University, Kayseri 38039, Turkey
66
Res. J. Appl. Sci. Eng. Technol., 4(1): 66-74, 2012
ERP implementation success respectively, in South
Africa, Malaysia and Bahrain. Ifinedo (2007) evaluate the
ERP system success by viewpoint of business and IT
managers in Finland and Estonia.
ERP systems have drawn much attention of many
researchers in the last two decades (Beheshti, 2006).
Many researches have studied on ERP implementation
success up to now. The first study on ERP systems’
critical success factors was studied by (Bingi et al., 1999).
Also, many researchers have studied (Holland and Light,
1999; Nah and Kuang, 2001; Hong and Kim, 2002;
Bradford and Florin, 2003; Somers and Nelson, 2004;
Soja, 2006; Ifinedo, 2007) ERP implementation success.
Researchers and practitioners have improved some
approaches to evaluate ERP system success. One of them
Diffusion of Innovation (DOI) that implies organization’s
innovation affect the diffusion of IT in organizations
(Bradford and Florin, 2003). Another approach is the
process approach that explains outcomes one event to
another by time sequence like in stages of implementation
life cycles (Somers and Nelson, 2004). Somers and
Nelson’s study (2004) indicates that success factors
influences on ERP systems can be temporal. CSFs relative
importance changes one stage to another in the
implementation life cycle (Somers and Nelson, 2004).
Other studies on ERP implementation success have been
described below.
In research, Parr et al. (1999) got ten experts who
participate in a total of 42 ERP implementation projects
opinion by the way of interview with them. Ten factors
necessary for successful implementation of ERP system
are identified. These factors were divided into four groups
related with management, personnel, software and project.
Holland and Light (1999) presented a number of
potential success factors which are divided into strategic
and tactical factors in ERP implementation. Nah and
Kuang (2001) research based on literature review yielded
a model of 11 critical success factors.
Al-Mashari et al. (2003) presented taxonomy of ERP
critical factors. A total of 12 factors were divided into
three dimensions related to ERP project stages, such as
setting-up, deployment and evaluation. Ifinedo (2007)
compared ERP success evaluations of business and IT
managers in his study. Data analysis of Ifinedo showed
that no significant statistical differences exist between the
two groups on the six dimensions (vendor/consultant
quality, system quality, information quality, individual
impact, work group impact, and organizational impact) of
ERP success operationalized with the exception of
vendor/consultant quality (Ifinedo, 2007).
There are many studies on ERP success in US and
west-Europe. On the other hand many researches are seen
in other countries. Some of them are listed below.
Sedera et al. (2002) examined ERP success across
different employment groups in Australian public
organizations. Zhang et al. (2003) examined critical
success factors of ERP implementation success in China.
Luliana (2006) studied CSF’s in Romanian SME’s ERP
implementation. Prokopiev et al. (2006), Jafari et al.
(2006) and Kamhawi (2008) examined factors affecting
Fuzzy cognitive mapping: A cognitive map that is a
qualitative model showing how a given system operates
is based on the defined variables and the casual relations
between these variables. Cognitive maps are especially
applicable and useful tools for modeling complex
relationships among variables. Cognitive mapping can be
used to examine the decision-makers’ maps, compare as
to their similarities and differences, and discuss them. In
addition the effects of different policy options can easily
be modeled (Özesmi and Özesmi, 2004). Cognitive maps
are directed graphs, or digraphs and their historical origins
are based on graph theory, which was formulized by
mathematician Euler in 1736 (Biggs et al., 1976). After
Euler, the graph theory has been extensively developed by
various mathematicians. Harary et al. (1965) included this
theory in numerical anthropology methods to study the
structural evaluation of empirical data and called it
digraph. Anthropologists have used signed digraphs to
represent differences between social structures in human
societies (Hage and Harary, 1983). Axelrod (1976)
transformed the graphics from the subjective observations
of anthropologists into to people’s perceptions graphics
and called it cognitive mapping (term first used by
Tolman, 1948). Many studies have used cognitive
mapping to examine and find out perceptions of complex
systems and to develop policies (Axelrod, 1976; Bauer,
1975; Brown, 1992; Carley and Palmquist, 1992; Cossette
and Audet, 1992; Klein and Cooper, 1982; Malone, 1975;
Montazemi and Conrath, 1986; Nakamura et al., 1982).
Kosko (1986) modified Axelrod’s binary cognitive
maps by applying fuzzy casual functions with real number
[-1, 1] to the connections. Thus, cognitive map has been
called as Fuzzy Cognitive Map (FCM). Kosko (1987) was
also the first to inference of FCM by using neural network
simulations, which ensure to model the effect of different
policy options.
FCM has been used as a tool to model a variety of
things in different fields, such as the physiology of
appetite (Taber and Siegel, 1987), political developments
(Taber, 1991), electrical circuits (Styblinski and Meyer,
1988), a virtual world of dolphins, shark and fish
(Dickerson and Kosko, 1994), organizational behavior
and job satisfaction (Craiger et al., 1996) and the
economic and demographic structures of world nations
(Schneider et al., 1998). In addition, many researchers
utilized FCM method in their studies for modeling of the
ecosystem (Özesmi, 1999), modeling of neuron fuzzy
cognitive maps (Tsadiras and Margaritis, 1997), using
fuzzy cognitive maps as a system model for effect
analysis (Peláez and Bowles, 1996), performance analysis
of personnel and modeling and analysis of business
67
Res. J. Appl. Sci. Eng. Technol., 4(1): 66-74, 2012
Fig. 1: An example of fuzzy cognitive mapping
performance assessment (Kardaras and Mentzas, 1997),
strategic planning (Tsadiras et al., 1995), increasing
effectiveness of information systems in strategic planning
process (Kardaras and Karakostas, 1999), the information
technology/information system investment evaluation
process in risky conditions to provide competitive
advantage for companies (Irani et al., 2002). Also, Yalç2n
and Seçme (2002) used cognitive mapping for
determining industrial problems and developing solutions
to these problems. Özesmi and Özesmi (2003) used FCM
to analyze the perceptions of different stakeholder groups
about a lake ecosystem in order to create a participatory
management plan. Çoban and Seçme (2005) used
cognitive mapping for the prediction of the effects of
privatization on the employees.
In this study, Fuzzy Cognitive Mapping (FCM)
method was used. Cognitive maps were obtained through
interviews with people who use ERP. Cognitive models
are complex systems, consisting of various variables have
interaction each other. Cognitive mapping is an effective
method because it reflects the subjective opinions of the
interviewee and it does not contain predetermined
questions. Cognitive mapping is a clear and transparent
method. It does not create anxiety on the part of the
interviewee and it is easy to prepare. It also helps to
collect data by give chance holistic comparison on the
subject and to set up cause-effect connections.
MATERIALS AND METHODS
Collecting data: In order to draw cognitive maps, the
questions of “What are the factors that affect ERP
application success?” and “How do these factors affect
each other?” were asked to the interviewees. Project
managers and plant managers, who applied ERP in their
plants, were interviewed. By using open ended questions,
the interviewee(s) were provided an atmosphere to
express their opinions freely without any restrictions and
direction on them. Respondents’ variables were listed on
a paper size of 50 cm × 35 cm. After the interviewee(s)
had made a list of all variables and were sure that they
had no more variables to add, they were asked about the
causal relationships among these variables indicating the
strengths of the relationship with a number between -1
and 1 [-strong (-1), -medium (-0,5), -a little (-0,25), zero
(0), a little (0,25), medium (0,5), strong (1)]. The
directions of the relationship were indicated with
arrowheads. As it can be understood from the casual
relationships, these are linguistic variables. Lingual
expressions like strong, medium and a little are regarded
as the natural representation of the preference or judgment
in this method. These characteristics indicate the
applicability of the fuzziness.
A total of 16 FCMs were used in the analyses. It was
seen that the average age of interviewees was 39,4 and the
drawing time of maps varied between 20 and 100 min.
One of the sample for cognitive maps is given in Fig. 1.
68
Res. J. Appl. Sci. Eng. Technol., 4(1): 66-74, 2012
The connection index means the density (D) of the
maps. To calculate density, the number of available
connections (C) is divided by the maximum number of
connections possible between N variables (Hage and
Harary, 1983). This shows how connected or sparse the
maps are:
20
The number of newly added variables
15
%95 confidence interval
10
D=
6
7
8
9
10
11
12
13
14
15
16
17
18
19
0
1
2
3
4
5
C
C
or D = 2
N ( N − 1)
N
(1)
-5
The type of variables in a map is important because
of the interaction between different types of variables. For
instance, same type of variables facilitates the
understanding of the structure of a cognitive map. In a
map, there exist three types of variables. These variables
are transmitter variables (forcing functions), receiver
variables (utility variables) and ordinary variables (means)
(Eden et al., 1992; Harary et al., 1965), which are defined
according to their out degree od(Vi) and in degree id(Vi).
Out degree is the row sum of absolute values of a
variable in the adjacency matrix. It shows the cumulative
strengths of connections exiting the variables, where N is
the total number of variables:
Fig. 2: The number of newly added variables and saturation
Saturation analysis: FCMs were created with different
individuals in such a way that the sample represents the
population sufficiently. In order to do this, Estimates 7.50
(Colwell, 1997) program was used to examine
accumulation curves of the total number of variables
versus number of interviews. Average accumulation curve
was created by using Monte Carlo techniques to randomly
select the order of the interviews many times, i.e, 500 and
to determine how the variables accumulate. In addition to
this, both the accumulated curves and the number of new
variables added per interview were evaluated in order to
obtain a sufficient sample.
At the end of the 16 interviews, a total of 51 different
variables were obtained from the maps. Beginning from
the 13th map, a clear slow down was observed in the
increase of the number of variable under 1. Beginning
from the 10th map, the number of new added variable
decreased less than 1, as it can be seen in Fig. 2.
This means that the sample size represents the
population sufficiently and the drawing maps are enough
for study. After this point, new added variables are
specific for the individual(s) and they indicate extreme
issues. In other words, the findings reflect that the
obtained sample size determined important wishes of
groups.
od (Vi ) =
N
∑
k −1
aik
(2)
In degree is the column sum of absolute values of a
variable. It shows the cumulative strength of variables
entering the variable.
id (Vi ) =
N
∑
k −1
(3)
aki
Transmitter variables have positive out degree, and
zero in degree. Receiver variables have a positive in
degree, and zero out degree. Ordinary variables can be
either receiver or transmitter variables, depending on the
ratio of their in degrees and out degrees.
The centrality (td (Vi)) of a variable is the sum of its
in degree and out degree and the centrality index shows
that the variable(s) is most mentioned.
Analysis of the data: There are two main ways of
analyzing FCMs. First way is to analyze the structural
indices values by using comparative statistical methods.
The second way is to use neural network simulations
found by Kosko (1992a, b). In this study, both analyses
were used for determining of the factors affecting ERP
application success.
After coding all the maps in Microsoft Excel as
square adjacency matrices, the indices values were
calculated, social cognitive maps were obtained, and these
maps were compared to each other by using statistical
techniques. The structural features of cognitive maps were
analyzed with Graph theory, which is the map in the form
of square adjacency matrices. Graph theory indices that
are density index, in degree, out degree, transmitter,
centrality, hierarchy index and complexity index were
calculated by using these matrices. In the following
equations, these mentioned indices are explained and each
one is showed with mathematical notations.
ci = td (V1 ) = od (V1 ) + id (V1 )
(4)
Another structural measure of a cognitive map is the
hierarchy index (h) (MacDonald, 1983):
h=
12
(n − 1)n(n + 1)
∑ [ od (v ) − (∑ od (v ))n]
i
i
69
i
2
(5)
Res. J. Appl. Sci. Eng. Technol., 4(1): 66-74, 2012
Table 1:
Graph theory indices for the individual cognitive maps &
social cognitive map
Individual
cognitive map
------------------Social
Mean
SD
cognitive map
No. of variables, N
20.58
1.62
51
No. of transmitter
2.92
2.39
3
Variables, T
No. of receiver variables R
2.92
2.11
0
No. of ordinary variables, O
14.75
3.60
48
Hierarchy index, h
0.08
0.04
0,36
No. of connections, C
48.5
011.74
4.8
Density, D
0.12
0.03
0.18
Connection/Variable, C/N
2.35
0.52
8.39
If h is equal to 1, the map is fully hierarchical; and if
h is equal to 0, the system is fully democratic. Democratic
maps are much more adaptable to local environmental
changes because of their high level of integration and
dependence (Sandell, 1996).
Thus the first step in analyzing cognitive maps is to
describe and tabulate the number of variables and
connections and the graph theory structural indices for
comparing among the interviewees or different groups of
stakeholders.
Individual cognitive maps of interviewees can be
augmented, and additively superimposed to form a Social
Cognitive Map (SCM). A social cognitive map displays
different behaviors rather than an individual cognitive
map. SCMs are also defined as “team maps” (Kosko,
1987; Eden et al., 1992; Kosko, 1992a; Kosko, 1992b;
Özesmi, 1999; Özesmi and Özesmi, 2003). In this paper,
social cognitive map refers to the augmented individual
map of project managers and plant managers, who applied
ERP in their plants in Kayseri.
The second method of computing FCMs is neural
networks simulations (Kosko, 1986; Kosko, 1987;
Reimann, 1998; Dickerson and Kosko, 1994). This
computational method is not necessarily concerned with
the structure, but it deals with the outcome, or inference
of the map (Kosko, 1987). In the neural network
computational method, a vector of initial states of
variables (In) is multiplied with the adjacency matrix A of
the cognitive map. The results are transformed to the
interval (0, 1) using a logistic function, which is 1/ (1+e"x). The transformation provides a better understanding
and representation of activation levels of variables and
enables us to compare qualitatively the causal output of
variables. These steps are repeated through the iterations
while the matrix values become steady state. To run
“what-if” scenarios, desired variables are set to a desired
value (0 or 1) at state vector for each simulation step and
the relative effects of desired variables are determined.
FCM simulation helps us to forecast the relationship
between the dependent variables and the independent
variables as the theoretically transformed results at the
end of the iterations can reach limit cycle or chaotic
attractor (Dickerson and Kosko, 1994). In this paper, our
matrix values became steady state before 20 iteration
steps and all values of variables obtained in simulations
go on a steady limit.
Therefore, to understand whether there is a
relationship between a number of described variables and
the interviewers’ age and the interview time or not,
correlation analyzes and t test were used. These analyses
showed that there is no relation between neither the
number of described variables and interviews’ age r
= -0.283; t = 0.000000), nor the number of described
variables and the interview time r = 0.308; t = 0.000003).
The mean age of our interviews was found 39.4±7.8. And
the mean time for drawing the maps is 47.13 min, while
the shortest interview time is 20 min and the longest
interview time is 100 min.
Graph theory indices values obtained from structural
analyses of cognitive maps are seen in Table 1. At the end
of the 16 cognitive maps, there are 20.58 variables in
average affecting ERP application success.
The number of receiver variables indicating the
expectations from the system was found as 2.92. The
number of transmitter variables, which affect the system
independently, was also found as 2.92. The ratio between
connections number per variable numbers; that is the
indicator of the density between the described variables
and the casual relations (or connection no per variable) is
2.35.
Hierarchy index showing the hierarchical structure in
the maps was calculated as 0.08. The closer to zero the
hierarchy index of a map, the more the map is democratic.
On the contrary the closer to one the hierarchy index of a
map, the more the map is hierarchical. The previous
studies showed that democratic maps can adapt to local
changes quiet fast because of their high level of
integration and dependence (Sandell, 1996). With respect
to individual cognitive maps, this can indicate a system
that is open to change. The average number of
connections (48.50) described in individual maps refer to
the collective dynamics of the system. It shows that the
interviewees think that the system is relatively complex
and versatile.
The findings of Social Cognitive Maps (SCM) were
analyzed and the analyses showed that a total of 51
variables were said by interviewees. While 3 of 51 were
ordinary variables, there were not any receiver variables.
RESULTS
The findings of structural analyses: Fuzzy cognitive
maps depend on the interview time or the length of the
subject, and also on the experience of the interview. Thus,
while comparing the variables and indices in cognitive
maps for different subjects and different interviews, maps
alone can not reflect the right results (Eden et al., 1992).
70
Res. J. Appl. Sci. Eng. Technol., 4(1): 66-74, 2012
Table 2: Some indices values in different studies to compare this study
No. of variables
No. of connections
Cossette and Audet (1992)
57.00
87.00
Özesmi (1999)
19.70
28.31
Yalç2n and Seçme (2002)
21.43
43.54
Özesmi and Özesmi (2004)
24.50
43.70
Çoban and Seçme (2005)
16.07
21.50
In this study
20.58
48.50
Connections/variable
1.53
1.64
2.07
1.78
1.32
2.35
Density
0.03
0.11
0.11
0.09
0.09
0.12
Table 3: The first ten most central variables and their centralities
Giving ERP education to team or last users
Conformity and reliability of data
Support of top level management
Right, fast and flexible reporting/output
Qualifications and experience of project manager
Responsible people from each departm. in project team
Knowing business and task of consultant
Collecting data fast and conveniently for the system
The flexibility of ERP system
The authority/efficiency of ERP consultant
Centrality
67.50
63.50
62.25
48.75
46.25
40.75
38.75
35.25
34.75
34.50
In degree
38.75
55.00
36.00
33.75
12.00
21.50
5.00
18.50
17.25
3.5.0
Out degree
28.75
85.00
26.25
15.00
34.25
19.25
33.75
16.75
17.50
31.00
to team and last users” is the most important variable and
it is under the effect of other variables (in degree).
The rest of the variables (48 variables) were the
transmitter variables. I n SCM, there were a total of 48
connections among variables. Connection number per
variable was 8.39. The density of SCM was 0.18 and
hierarchy index was 0.36. This system revealed by our
research is a rather democratic and low density system.
Though project managers and plant managers, who
applied ERP in their plants, contacted different points, as
a whole they all indicated the similar model.
When the indices values of this study are compared
with other studies, which previously used the same
method, the model in this study appears to be denser and
to have a stronger interaction (Table 2).
Especially the indices number of connections and the
connection number per variable are higher than other
studies. This situation shows that variables in the system
have more dense interactions together. Same situation can
also be said for density indices. The power of the
interaction of the system shows the effect of the density
indices value. The more dense the maps, the more
complex subjects and power interactions between
variables in maps.
While the variables were ordered according to their
centrality degree, the most central (most important)
variable was “to give ERP education to team or last
users”. The second most central variable was “conformity
and reliability of data” (Table 3).
We can understand how much these central variables
affect and are affected by other variables by looking at the
out degree and in degree. Hence, it is easy to see and
understand the relationships between variables in the
maps. If a variable has a big effect on other variables, this
is out degree. On the contrary, if a variable is affected by
the other variables, this is in degree. When we look at the
Table 3, we can see that the “qualifications and
experience of the other project manager” has big effect on
other variables (out degree) and, “to give ERP education
Findings of neural networks simulations: The second
way of analyzing FCM is the neural network simulations.
After the cognitive maps are drawn and the adjacency
matrix is coded, it is possible to run the model in order to
see where the system will go and to determine the
system’s steady state. In the simulation analyses, we have
used logistic function (1(1+e-1*x) to transform the results
into interval [0, 1]. This non-negative transformation
allows for a better understanding and representation of
activation levels of variables, and also enables a
qualitative comparison among the casual output of
variables. The resulting transformed vector is repeatedly
multiplied by the adjacency matrix and transformation is
carried out until the system converges into a fixed point.
In our study, the system converges in less than 20
simulation time steps.
In this study, two scenarios were run to predict the
effects of variables on ERP success: the first one is the
analysis of the factors, which will meet the expectations
from the ERP systems of business and raise the success of
system, and the second is to predict these variables’
effects on other variables.
According to the first scenario, if ERP education
variable is enabled, expanded or raised the efficiency,
which variables raise because of the effect of education.
Because of this, the most central variable, which is
“giving ERP education to team and last users”, is
strengthened. At the end of this scenario, some variables
became more important than the other variables. These
are; meeting the business needs of ERP system, the
efficiency of project plan, project management and time
schedule, system authority of last users, the existence of
responsible persons from each department in the project
team, the flexibility of ERP system and collecting the data
fast and conventional to the system (Fig. 3).
71
Res. J. Appl. Sci. Eng. Technol., 4(1): 66-74, 2012
ERP system implementation will cause to waste all
resources spend for this project. In this study fuzzy
cognitive mapping method was used to determine the
factors affecting ERP application success. Cognitive maps
were obtained through interviews with project managers
and plant managers who applied ERP system in their
plant. This is the first attempt to determine the factors
affecting ERP success using cognitive mapping.
A total of 16 FCMs were used in the analyses. Fuzzy
cognitive mappings were analyzed in two ways. In the
first way, the structural indices values were analyzed by
using comparative statistical methods. In the second way,
neural network simulations were used.
According to the results of structural analyses of 16
cognitive maps, a total of 51 variables were said by
interviewers in social cognitive maps. 3 of 51 variables
were found ordinary variables. The rest of the variables
(48 variables) were the transmitter variables and there was
not found any receiver variables. In SCM, there were a
total of 48 connections among variables. Connection
number per variable was 8.39. This system revealed by
our research is a rather democratic and low density
system.
When the indices values of this study are compared
with other studies, which previously used the same
method, the model in this study appears to be denser and
have a stronger interaction. Especially the indices number
of connections and the connection number per variable
are higher than other studies. This situation shows that
variables in the system have more dense interactions
together. Same situation can also be said for density
indices.
While the variables were ordered according to their
centrality degree, the most central (most important)
variable was “to give ERP education to team and last
users”. The second most central variable was “conformity
and reliability of data”.
We saw that the “qualifications and experience of the
other project manager” has big effect on other variables
(out degree) and, “to give ERP education to team and last
users” is the most important variable and it is under the
effect of other variables (in degree).
In this study, two scenarios were run to predict. The
first one is the analysis of the factors, which will meet the
expectations from the ERP systems of business and raise
the success of system, and the second is to predict these
variables’ effects on other variables.
According to the first scenario, if ERP education
variable is enabled, expanded or raised the efficiency,
which variables raise by the effect of education. Because
of this, the most central variable, which is “giving ERP
education to team and last users”, is strengthened. At the
end of this scenario, some variables became more
important than the other variables. These are; meeting the
business needs of ERP system, the efficiency of project
plan, project management and time schedule, system
authority of last users, the existence of responsible
0.12
1.00
0.08
0.06
0.04
0.02
The flexibility of ERP system
Responsible persons
should be from each
department in project
team
The integration capability
with business supplier and
customers
Self-sacrifice of
project team
Giving ERP education to
team and last user
Meet the business needs
of ERP system
0
Fig. 3: The relative effect of giving ERP education to team and
last users
0.12
1.00
0.08
0.06
0.04
0.02
The flexibility of ERP system
Responsible persons
should be from each
department in project
team
The integration capability
with business supplier and
customers
Self-sacrifice of
project team
Giving ERP education to
team and last user
Meet the business needs
of ERP system
0
Fig. 4: The relative effect of support of top level manager
According to the second scenario, if top level
managers support the use of ERP system, this can have
important implications for the business. In order to see
this interaction, the variable indicating “the support of top
level manager” is strengthened or extended. At the end of
the second scenario, the following variables are shown:
meeting the business needs of ERP system, giving ERP
education to team and last users, self-sacrifice of project
team, existence responsible persons from each department
in the project team, the integration capability with
business suppliers and customers and the flexibility of
ERP system (Fig. 4).
DISCUSSION
ERP system implementation projects need to spend
a large amount of money, time and efforts. The failure of
72
Res. J. Appl. Sci. Eng. Technol., 4(1): 66-74, 2012
persons from each department in the project team, the
flexibility of ERP system and collecting the data fast and
conventional to the system.
According to the second scenario, if top level
managers support the use of ERP system, this can have
important implications for the business. In order to see
this interaction, the variable indicating “the support of top
level manager” is strengthened or extended. At the end of
the second scenario, the following variables are shown:
meeting the business needs of ERP system, giving ERP
education to team and last users, self-sacrifice of project
team, existence responsible persons from each department
in the project team, the integration capability with
business suppliers and customers and the flexibility of
ERP system.
Briefly, it is appeared that most important critical
success factors at ERP implementation stage are training
the ERP team and last users, and Top management
support.
Çoban, O. and G. Seçme, 2005. Prediction of socioeconomical consequences of privatization at the firm
level with fuzzy cognitive mapping. Inf. Sci., 169:
131-154.
Colwell, R.K., 1997. Estimate S: Statistical Estimation of
Species Richness and Shared Species from Samples.
Version 5 User’s Guide and Application Published,
Retrieved from: http: //www. viceroy. eeb. uconn.
edu/ estim ates.
Cossette, P. and M. Audet, 1992. Mapping of an
idiosyncratic schema. J. Man. St., 29: 325-347.
Craiger, J.P., R.J. Weiss, D.F. Goodman and A.A. Butler,
1996. Simulating organizational behavior with fuzzy
cognitive maps. Int. J. Computational Intelligence
Organizations, 1: 120-133.
Dickerson, J.A. and B. Kosko, 1994. Virtual worlds as
fuzzy cognitive maps. Presence, 3: 173-189.
Eden, C., F. Ackerman and S. Cropper, 1992. The
analysis of cause maps. J. Man. St., 29: 309-323.
Hage, P. and F. Harary, 1983. Structural Models in
Anthropology. Oxford University Press, New York.
Harary, F., R.Z. Norman and D. Cartwright, 1965.
Structural Models: An Introduction to the Theory of
Directed Graphs. John Wiley & Sons, New York.
Holland, C., and B. Light, 1999. A Critical Success
Factors Model for ERP Implementation. IEEE
Software, May/June, 30-35.
Hong, K.K. and Y.G. Kim, 2002. The critical success
factors for ERP implementation: An organizational
fit perspective. Inf. Manage., 40(1): 25-40.
Ifinedo, P., 2007. An empirical study of ERP success
evaluations by business and IT managers. Inf. Man.
Comp. Security, 15(4): 270-282.
Irani, Z., A. Sharif, E.D. Love and C. Kahraman, 2002.
Applying concepts of fuzzy cognitive mapping to
model: The IT/IS investment, evaluation process.
Inter. J. Prod. Econ., 75: 199-211.
Jafari, S.M., M.R. Osman, R.M. Yusuff and S.H. Tang,
2006. ERP systems implementation in Malaysia: The
importance of critical success factors. Int. J. Eng.
Tech., 3(1): 125-131.
Kamhawi, E.M., 2008. Examining the Factors Affecting
Project andBusiness Success of ERP Implementation.
In: Gunasekaran, A., (Ed.), Global Implications of
Modern Enterprise Information System. Information
Science Reference-IGI, PA, USA, pp: 1-18.
Kardaras, D. and G. Mentzas, 1997. Using Fuzzy
Cognitive Maps to Model and Analyse Business
Performance Assessment. Proceedings of the 2nd
Annual International Conference on Industrial
Engineering Applications and Practice, pp: 63-68.
Kardaras, D. and B. Karakostas, 1999. The use of fuzzy
cognitive maps to simulate the information systems
strategic planning process. Inf. Software Technol.,
41(4): 197-210.
REFERENCES
Al-Mashari, M., A. Al-Mudimigh and M. Zairi, 2003.
Enterprise resource planning: A taxonomy of critical
factors. Eur. J. Op. Res., 146: 352-364.
Axelrod, R., 1976. Structure of Decision: The Cognitive
Maps of Political Elites. Princeton University Press,
Princeton, NJ.
Bauer, V., 1975. Simulation, Evaluation and Conflict
Analysis in Urban Planning, In: Baldwin, M.M.,
(Ed.), Portraits of Complexity, Applications of
Systems Methodologies to Societal Problems.
Battelle Institute, Columbus, OH, pp: 179-192.
Beheshti, H.M., 2006. What managers should know about
ERP/ERP II. Man. Res. News, 29(4): 184-193.
Biggs, N.L., E.K. Lloyd and R.J. Wilson, 1976. Graph
Theory. 17Clarendon Press, Oxford, pp: 36-1936.
Bingi, P., M. Sharma and J. Golda, 1999. Critical issues
affecting an ERP implementation, Inf. Syst. Man.,
16(3): 7-14.
Bradford, M. and J. Florin, 2003. Examining the role of
innovation diffusion factors on the implementation
success of ERP systems. Int. J. Account. Inf. Syst.,
4(3): 205-225.
Brown, S.M., 1992. Cognitive mapping and repertory
grids for qualitative survey research: Some
comparative observations. J. Manage. Stud., 29: 287307.
Calisir, F. and F. Calisir, 2004. The relation of interface
usability characteristics, perceived usefulness and
perceived ease of use to end-user satisfaction with
ERP systems. Comput. Hum. Behav,. 20(3): 505515.
Carley, K. and M. Palmquist, 1992. Extracting,
representing and analyzing mental models. Soc.
Forces, 70: 601-636.
73
Res. J. Appl. Sci. Eng. Technol., 4(1): 66-74, 2012
Prokopiev, S., L. Seymour and J.P. Van Belle, 2006. An
Emp. Inves. of Factors Affecting ERP Impact. In:
Khalid, S.S., (Ed.), Man. Inf. in The Digital
Economy: Issues & Solutions, Proc. of the 6th
Int.Bus. Inf. Man. Ass. Con., 415-423, 19-21, Bonn,
Germany.
Reimann, S., 1998. On the des. of art. auto-ass. neural
networks. Neural Networks, 11: 611-621.
Sandell, K., 1996. Sus. in theory and practice: a
conceptual framework of eco-strategies and a case
study of low-resource agriculture in the dry zone of
Sri lanka. In: Hjort-af-Ornãs, A. (Ed.), Appr. Nature
from Local Communities: Security Per. and
Achieved. Linköping University, Sweden, pp:163197.
Schneider, M., E. Shnaider, A. Kandel, G. Chew, 1998.
Automatic construction of FCMs. Fuzzy Sets Syst.,
9: 161-172.
Sedera, D., G. Gable and A. Palmer, 2002. ERP systems
impacts: A Delphi study of Australian public sector
organizations, Pro. of The Pacific Asia Con. on Inf.
Systems, Tokyo, Japan, pp: 584-600.
Soja, P., 2006. Success factors in ERP systems
implementations: Lessons from practice. J. Enterprise
Inf. Manage., 19(6): 646-661.
Somers, T. and K. Nelson, 2004. A taxonomy of players
activities across the ERP project life cycle. Inf.
Manage., 41(3): 257-278.
Styblinski, M.A. and B.D. Meyer, 1988. Fuzzy Cognitive
Maps, Signal Flow Graphs and Qualitative Circuit
Analysis. Proceeding. of the 2nd IEEE International
Conference on Neural Networks, San Diego, CA,
pp: 549-556.
Taber, W.R., 1991. Knowledge processing with fuzzy
cognitive maps. Expert Syst. App., 2: 83-87.
Taber, W.R. and M.A. Siegel, 1987. Estimation of expert
weights using fuzzy cognitive maps. Pro. of the 1st
IEEE International Conference on Neural Net., pp:
319-325.
Tolman, E.C., 1948. Cognitive maps in rats and men.
Psychol. Rev., 55(4): 189-208.
Tsadiras, A.K. and K.G. Margaritis, 1997. Cognitive
mapping and certainty neuron fuzzy cognitive maps.
Inf. Sci., 101: 109-130.
Tsadiras, A., K. Margaritis and B. Mertzios, 1995.
Strategic planning using extended fuzzy cognitive
maps. St. Inf. Con, 4(3): 237-245.
Yalç2n, N. and G. Seçme, 2002. Kayseri Sanayinin
Problemleri ve Çözüm Önerileri: Bulan2k BiliÕsel
Haritalama Örne—i, II. Orta Anadolu Kongresi, Ni—de
Üniversitesi, Ni—de, Türkiye.
Zhang, L., M.K.O. Lee, Z. Zhang and P. Banerjee, 2003.
Crit. success factors of ERP systems implementation
success in China, Pro of the 36th Hawaii International
Conference on Sys. Sci, pp: 23.
Klein, J.H. and D.F. Cooper, 1982. Cognitive maps of
decision-makers in a complex game. J. Operational
Res. Soc., 33: 63-71.
Kosko, B., 1986. Fuzzy cognitive maps. Int. J. ManMachine Stud., 1: 65-75.
Kosko, B., 1987. Adaptive inference in fuzzy knowledge
networks, In: Proceedings of the First IEEE
International Conference on Neural Networks
(ICNN-86), San Diego, CA, pp: 261-268.
Kosko, B., 1992a. Fuzzy Associative Memory Systems.
In: Kandel, A., (Ed.), Fuzzy Expert Systems, CRC
Press, Boca Raton, pp: 135-162.
Kosko, B., 1992b. Neural Networks and Fuzzy Systems:
A Dynamical Systems Approach to Machine
Intelligence, Prentice-Hall, Englewood Cliffs, NJ.
Luliana, S., 2006. Critical success factors in Romanian
SME’s ERP implementation. Proceedings of the 3rd
International Workshop IE&SI, pp: 308-315,
Timisoara, Romania.
MacDonald, N., 1983. Trees and Networks in Biological
Models. John Wiley and Sons, NY.
Malone, D.W., 1975. An introduction to the Application
of Interpretive Structural Modeling, In: Baldwin,
M.M., (Ed.), Portraits of Complexity: App. of
Systems Methodologies to Societal Problems.
Battelle Institute, Columbus, OH, pp: 119-126.
Montazemi, A.R. and D.W. Conrath, 1986. The use of
cognitive mapping for information requirements
analysis. MIS Quarterly, pp: 10: 45-55.
Nah, F. and J. Kuang, 2001. Critical factors for successful
implementation of enterprise systems. Business
Process Man J., 7(3): 285-296.
Nakamura, K., S. Iwai and T. Sawaragi, 1982. Decision
support using causation knowledge base. IEEE Trans.
Syst. Man. Cybernet. SMC, 12: 765-777.
Özesmi, U., 1999. Modeling ecosystems from local
perspectives: Fuzzy cognitive maps of the Kizilirmak
Delta wetlands in Turkey. Proceedings of 1999
World Conference on Natural Resource Modelling,
Halifax, NS, Canada, pp: 23-25.
Özesmi, U. and S. Özesmi, 2003. A participatory
approach to ecosystem conservation: Fuzzy cognitive
maps and stakeholder group analysis in Uluabat
Lake. Turkey Env. Man., 31(4): 518-531.
Özesmi, U. and S.L. Özesmi, 2004. Eco. models based on
people’s knowledge: A multi-step fuzzy cognitive
mapping approach, Ec. Mod., 176(1-2): 43-64.
Parr., A., G. Shanks and P. Darke, 1999. Idenfication of
necessary factors for successful implementation of
ERP systems, New Information Technologies in
Organizational Processes-Field Studies and
Theoretical Reflections on the Future of Work,
Kluwer Academic Pub., NY, pp: 99-119.
Peláez, C.E. and J.B. Bowles, 1996. Using fuzzy cog
maps as a system model for failure modes and effects
analysis. Inf. Sci., 88(1-4): 177-199.
74